Rac function is crucial for cell migration but is not required for spreading and focal adhesion formation

Abstract

Cell migration is commonly accompanied by protrusion of membrane ruffles and lamellipodia. In two-dimensional migration, protrusion of these thin sheets of cytoplasm is considered relevant to both exploration of new space and initiation of nascent adhesion to the substratum. Lamellipodium formation can be potently stimulated by Rho GTPases of the Rac subfamily, but also by RhoG or Cdc42. Here we describe viable fibroblast cell lines genetically deficient for Rac1 that lack detectable levels of Rac2 and Rac3. Rac-deficient cells were devoid of apparent lamellipodia, but these structures were restored by expression of either Rac subfamily member, but not by Cdc42 or RhoG. Cells deficient in Rac showed strong reduction in wound closure and random cell migration and a notable loss of sensitivity to a chemotactic gradient. Despite these defects, Rac-deficient cells were able to spread, formed filopodia and established focal adhesions. Spreading in these cells was achieved by the extension of filopodia followed by the advancement of cytoplasmic veils between them. The number and size of focal adhesions as well as their intensity were largely unaffected by genetic removal of Rac1. However, Rac deficiency increased the mobility of different components in focal adhesions, potentially explaining how Rac - although not essential - can contribute to focal adhesion assembly. Together, our data demonstrate that Rac signaling is essential for lamellipodium protrusion and for efficient cell migration, but not for spreading or filopodium formation. Our findings also suggest that Rac GTPases are crucial to the establishment or maintenance of polarity in chemotactic migration.

Keywords: Actin; Adhesion; CAAX; Chemotaxis; Filopodia; Lamellipodia; Migration; Rac1.

Fig. 1.

Rac1-deficient MEFs are unable to form lamellipodia and ruffles. (A) Genotyping of Rac1^fl/fl cells and individual Rac1^−/− clones for genotypes as indicated. (B) Western blot of Rac1^fl/fl and individual Rac1^−/− clones with Rac1/3 antibody (upper panel) and tubulin antibody (lower panel). (C,D) Phalloidin staining of Rac1^fl/fl (C) and Rac1^−/− MEFs (D). (E–L) Growth factor stimulation. Rac1^fl/fl MEFs (E–H) and Rac1^−/− MEFs (I–L) were either starved (E,I), or stimulated with PDGF (F,J), EGF (G,K) or HGF (H,L). (M) Quantification of growth factor stimulations. Data were collected from three independent experiments and are expressed as means ± s.e.m.; n = total number of cells analyzed.

Fig. 2.

Rac1, Rac2 and Rac3 restore lamellipodia and interact with the WAVE complex, but not RhoG and Cdc42. (A–F′) Expression of constitutively active Rho GTPases in Rac1^fl/fl and Rac1^−/− cells. Rac1^fl/fl (A,A′,C,C′,E,E′) and Rac1^−/− (B,B′,D,D′,F,F′) MEFs were transfected with myc-tagged Rac1-L61 (A–B′), RhoG-L61 (C–D′) and Cdc42-L61 (E–F′), fixed and stained with phalloidin (A,B,C,D) and anti-myc (A′,B′,C′,D′). Note the absence of lamellipodia in D and E as opposed to A, B, C and E. (G) Quantification of phenotypes caused by expression of different Rho GTPases. Rac1^fl/fl and Rac1^−/− MEFs without overexpression of any Rho GTPase (–) or transfected with myc-tagged versions of constitutively active (L61) GTPases as indicated were evaluated. Cells were first scored for the presence of lamellipodia (blue). In case no lamellipodia were present, the presence of filopodia was assessed (light orange), and in case neither lamellipodia nor filopodia were detectable, cells were scored as ‘no protrusion’ (purple). Data were collected from three independent experiments and are expressed as means ± s.e.m.; n = total number of cells analyzed. (H) WAVE complex interacts with Rac1, Rac2 and Rac3, but not RhoG and Cdc42. Pull down of recombinantly expressed Rho GTPases and controls as indicated shows Sra-1 binding to Rac1, Rac2 and Rac3 (upper panel). The lower panel shows input of recombinant proteins. As GST–RhoG-L61 was proteolytically cleaved in part (asterisk), MBP-tagged RhoG-V12 was used as confirmation.

Fig. 3.

Rac1-ΔCAAX is capable of robust but not full rescue of lamellipodium formation. Rac1^fl/fl (A) and Rac1^−/− (B,C) cells expressing myc-Rac1-L61-ΔCAAX form lamellipodia with distinct phenotypes. For quantification of robustness of lamellipodium formation, categories were assigned according to their contours as follows: regular lamellipodia associated with convex cell outline (see A″), regular lamellipodia associated with concave cell outline (B″) and small, irregular lamellipodia (C″). Cells were transfected with myc-Rac1-L61 (not shown, see also Fig. 2) and myc-Rac1-L61-ΔCAAX and stained with phalloidin (A,B,C) and anti-myc (A′,B′,C′) to identify transfected cells. Enlargements of the actin cytoskeleton at the cell periphery of regions boxed in A–C are shown in A″,B″,C″. Scale bars in C′ (for A,A′,B,B′,C) and C″ (A″,B″). (D) Quantification of lamellipodial phenotype. Rac1^fl/fl and Rac1^−/− MEFs expressing Rac1 constructs as indicated were assessed for the presence of convex, concave or irregular lamellipodia. Data are from three independent experiments and expressed as means ± s.e.m.; n = total number of cells analyzed. (E) Rac1^fl/fl and Rac1^−/− cells transfected with myc-tagged Rac1 constructs, as indicated, were separated into cytosol (C) and membrane (M) fractions. Samples were probed with antibodies as indicated. Anti-myc-probed membrane fractions are additionally shown with contrast enhancement (bottom panels).

Fig. 4.

Migration capacity is strongly reduced in Rac1-deficient MEFs. (A) Selected frames from wound healing movies of control (Rac1^fl/fl) and Rac1^−/− cells. (B) Average wound closure for each cell type over time. Rac1^−/− cells are not able to close the wound after 20 hours (see supplementary material Fig. S5). Data were collected from three independent experiments; n = total number of movies analyzed. Error bars indicate ± s.e.m. (C) Average wound closure for each cell type treated with Rho kinase inhibitor Y27632 over time. Data were collected from two independent experiments; n = total number of movies analyzed. Error bars represent ± s.e.m. (D) Wound closing speed of Rac1^fl/fl and Rac1^−/− cells with or without Y27632 treatment. Note that Rho kinase inhibition increases the wound closure speed in both cell types to the same extent. Error bars indicate ± s.e.m. (E–J) The leading front of Rac1^−/− cells facing the wound area is devoid of lamellipodia but has numerous filopodia. Rac1^fl/fl (E–E″,G–G″,I) and Rac1^−/− cells (F–F″,H–H″,J) 4 hours after wounding were stained with phalloidin (E′,F′,G′,H′) and anti-Abi (E″,F″,G″,H″). (E,F,G,H) phase contrast, (I,J) merged images.

Fig. 5.

Random migration and chemotaxis are abrogated in Rac1-deficient MEFs. (A) Random migration of Rac1^fl/fl, Rac1^−/− and Rac1^−/− cells expressing dominant-negative Rac1-N17 (medians 0.380, 0.090, 0.000, respectively). Data sets are from three independent experiments, n = total number of cells analyzed. (B) Chemotactic migration of Rac1^fl/fl and Rac1^−/− cells towards a gradient of 2.5% serum and 100 ng/ml HGF (medians 0.310 and 0.040, respectively). Data sets are from three independent experiments, n = total number of cells analyzed. (C) Trajectory plots of Rac1^fl/fl and Rac1^−/− cells in the chemotaxis assay show all individual cell paths. Migration paths towards the HGF-gradient are colored black, migration paths away from the growth factor source are colored red. (D) Rose plots of 10° segments showing the frequency of a given direction of migration paths during chemotaxis. (E) Forward migration index (FMI) of Rac1^fl/fl and Rac1^−/− cells during chemotaxis (medians 0.370 and 0.070, respectively). Box and whiskers plots show medians, 10th, 25th, 75th and 90th percentiles and dots show individual data points.

Fig. 6.

Rac1 is not required for cell spreading. (A–F′) Immunofluorescence of spreading cells. Rac1^fl/fl (A–C′) and Rac1^−/− cells (D–F′) were stained with phalloidin (A,B,C,D,E,F) and anti-Abi (A′,B′,C′,D′,E′,F′) 15 minutes (A,A′,D,D′), 60 minutes (B,B′,E,E″) and 24 hours (C,C′,F,F″) after plating on 25 µg/ml fibronectin. Note that localization of Abi at the lamellipodium tip, prominent in B′ and C′, is absent in Rac1^−/− cells. (G–J) Quantification of the spreading area on different fibronectin concentrations or extracellular matrices. Rac1^fl/fl and Rac1^−/− MEFs were plated on 5 µg/ml fibronectin (G), 25 µg/ml fibronectin (H), 25 µg/ml laminin (I) or 0.2% gelatin (J) as detailed in the Materials and Methods. All datasets are from three independent experiments. Error bars indicate the s.e.m.; n = total number of cells analyzed.

Fig. 7.

Rac1 is not essential for focal adhesion and filopodium formation. (A–F) Vinculin staining of Rac1^fl/fl (A–C) and Rac1^−/− (D–F) cells 15 minutes (A,D), 60 minutes (B,E) and 24 hours (C,F) after plating. After 24 hours, the distribution of focal adhesions is more variable between cells than at earlier time points. Some cells show focal adhesions evenly distributed under the cell (yellow arrowheads in C), while some display focal adhesions mainly at peripheral attachment points (blue arrowheads in C). As opposed to controls after 24 hours of spreading, most peripheral attachment points in Rac1^−/− cells appeared equidistant from the nucleus, indicating lack of polarization (yellow arrows in F). (G,H) Rac1^−/− MEFs spread by employing filopodia. Selected frames of phase-contrast movies of Rac1^fl/fl (G) and Rac1^−/− cells (H) acquired during spreading. Arrowheads point to lamellipodia, arrows point to filopodia. (I–K) Focal adhesions are formed at the base of filopodia in Rac1^−/− cells. Selected phase-contrast (I) and green epifluorescence channel (J) frames of Rac1^−/− cells expressing GFP–VASP, show accumulation of VASP in focal adhesions formed at the base of protruding filopodia. (K) shows merged phase contrast (gray) and GFP–VASP (green) frames. (L) Maturation of focal adhesions over time is shown in a merge of GFP-VASP images from three different time points. Blue corresponds to GFP-VASP localization after 12 minutes, green corresponds to 16 minutes and red to 20 minutes.

Fig. 8.

Analysis of focal adhesion parameters in Rac1^−/− MEFs. (A–F) Assessment of focal adhesion parameters from vinculin stainings. Representative images of Rac1^fl/fl (A,A′) and Rac1^−/− (B,B′) cells stained with vinculin antibodies (A,B). Images were thresholded as detailed in Materials and Methods; a binarized threshold (A′,B′) is shown to represent data extraction. From these images, number of adhesions per cell (C), number of adhesions per cell area (D), focal adhesion intensity (E) and focal adhesion size (F) were calculated. 30 cells were analyzed for each cell type.

Fig. 9.

Turnover of focal adhesion components. (A–D) FRAP analysis of EGFP–zyxin in Rac1^fl/fl (A) and Rac1^−/− (B) MEFs. Representative frames of bleaching experiments show EGFP–zyxin accumulation in focal adhesions before (pre FRAP, left panel), immediately after bleaching (0 seconds) and during fluorescence recovery at the time points indicated. Yellow quadrilaterals mark bleached areas. (C–H) Recovery curves of normalized fluorescence intensities of EGFP-tagged zyxin (C,D), paxillin (E,F) and VASP (G,H) in Rac1^fl/fl (blue) and Rac1^−/− cells (orange). C,E,G show arithmetic means with s.e.m. for acquired time points before and after bleaching. MF, mobile fraction, IF, immobile fraction. Mobile fractions are shown as percentages of the total fraction (sum of mobile and immobile fraction). D,F,H show fitted curves of averaged data from which half times of recovery (t_1/2) were calculated. N = number of analyzed movies; respective equations of curve fits for each component are displayed in the figure, equation coefficients are given in supplementary material Table S3.